Generates images from text prompts using the Kandinsky neural network API, with support for styles, negative prompts, and configurable image dimensions.
